Colour and B&W Digital Infrared

Link to the image gallery with twelve IR image pairs: Colour and B&W Digital Infrared Gallery

Since the late 90’s I have always had some sort of digital camera that allowed me to do infrared (IR) photography. Back in the day, some point & shoot cameras had weak internal IR blocking filters and with the addition of a visible light blocking filter that allows IR to pass, like a Hoya R72 (720nm band-pass) for example, one could just manage to do daytime IR photos. I used a number of different cameras; an Olympus C4040, a Nikon CoolPix 990, a Nikon CoolPix 5400 and lastly the original Fujifilm X100.

However, modifying a camera to remove the IR blocking filter is a much better option and will yield more dramatic results and much faster shutter speeds to allow for the use of lower ISOs. My first modified camera was a Canon EOS 20D digital SLR. For some samples taken with that camera way back in 2006 on a trip to Oregon, see this gallery: Infrared with a Canon EOS-20D. The issue with older DSLRs is that not only do they have an AF sensor that is separate from the image sensor, but they also do not have any AF micro adjustment control which would allow you to calibrate the IR offset for specific lenses. The reason that’s an issue is that the AF sensor uses visible light to focus, but most lenses have a focus shift when imaging in IR, so you would get blurry shots. The EOS 20D was calibrated at the place that did the IR conversion (more on that in a bit) using a mid-range zoom, but for example, when I put on my EF 400mm f/5.6L, it was way off. With practice, I got good at letting the camera AF, then twist the focusing ring ever so slightly to compensate for the IR offset. Newer DSLRs have AF micro adjustments than can be done on a lens by lens basis, so you could likely dial things in much better. However for the ultimate in digital IR, a mirrorless camera is the way to go.

There are two significant advantages to using a mirrorless camera for digital IR. Firstly, since a mirrorless camera feeds a live image from the sensor to your EVF or back LCD, you are seeing the IR effect while composing the shot. A converted DSLR still shows a regular, through-the-lens normal colour view of your scene, so until you take a shot and hit playback, you’ll have no idea how dramatic the IR effect might even be, nor even if the camera is giving you a remotely accurate exposure. Secondly, since the main image sensor is used for metering and autofocus, you don’t have to worry about unpredictable exposures or blurry images due to any IR focus shift.

After the EOS-20D, I had a Panasonic GF-1 converted to IR and realized just how much better a mirrorless camera was for digital IR work. A few years after, I had sold off my Panasonic gear and mostly transitioned to my Fujifilm system and when I got my X-Pro2, I decided to hang onto my X-E2 and have it IR converted, and that is still the camera I am using to this day for digital IR. I had those three cameras converted by LifePixel, just outside of Seattle, Washington, and they have always done a superb job. I have sent numerous customers their way as well and as far as I know, everyone has been very happy with their conversions too.

If you don’t have a spare camera from an upgrade that you can use for an IR conversion, a good option might be to buy a second hand mirrorless camera. For example, right now, we have a used Fujifilm X-T3 for $1,100, and for DSLRs, we have a range of Canon and Nikon models for $300 to $900. Contact Ken here, who handles used cameras, and he can let you know what we might have that would be suitable for an IR conversion.

There are many different filter choices when you have a camera converted, from deep IR (830 nm bandpass) where you will basically only get a B&W image out of camera, to a so-called “Hyper Color IR” (470nm) which lets through a lot of colour light and doesn’t block any IR. One of the most common standard IR filters is at 720nm and it offers just the slightest hint of colour with a captured image but still demonstrates great contrast and IR effect. I chose to go just a little “lighter” with my X-E2 and got the “Enhanced Color IR” filter with a 665nm band-pass. It lets through just a little more colour than the 720nm filter while still maintaining excellent contrast for IR, giving you dark skies, bright foliage and so on. With appropriate processing, one can get both interesting colour IR effects and high contrast B&W as well. The image gallery linked to above has twelve images, each processed as a colour IR and as a B&W IR.

There are many different ways of getting an interesting colour IR and you can easily skew the colours one way or the other. I will detail what I did for the images in the gallery, but feel free to play with things to achieve your own look. For an early look at my first tests with the converted X-E2, see the following posting from 2016 on my personal blog. Note that I used PhotoNinja for my initial raw conversions there, but these days I use Capture One Pro since it also has a wider adjustment range for WB when compared to Photoshop or Lightroom: http://www.mikemander.com/2016/08/infrared-with-fujifilm-x-e2.html

Here is some detail to the post-processing steps I used for the images in the Beau Photo gallery…

Firstly, I used Capture One Pro and dramatically skewed the white balance to 1,580K and -33 on the green/magenta tint. Note that Adobe raw conversions are limited to 2,000K as the lowest WB. Not that you couldn’t make a 2000K setting work, but I do find it easier to create the “blue sky” channel swap images with a lower WB setting. Out of camera, the raw files in Capture One looked like this…

After the dramatic WB adjustment, we get the following, and that’s what I send from Capture One to Photoshop as a 16-bit PSD file…

From there, in Photoshop I apply a channel-swap using the following settings (all done in one step but here showing the three different channel adjustments by selecting the Output Channel). By the way, you can click on any of the images in the posting to open them larger in a new window (except the headline image)…

In Photoshop’s Channel Mixer, the red channel gets 0% red and green and 100% blue, green stays the same, and blue gets 0% blue and green and 100% red. Those settings effectively swap the red and blue channels of the image, resulting in this where the red skies are now blue…

Then, I add contrast, clarity and saturation, as well as skew the colours slightly, for example skewing the blues away from magenta a little, resulting in this…

One can do a lot more with the colour, selectively making adjustments and creating a greater variety of hues, but I’ve kept it quite simple here. Then it gets more interesting…

One thing I’ve always loved about Kodak’s B&W HIE high speed infrared film (sadly, long discontinued) is the lack of an anti-halation backing layer. What happens is that the brightest IR light shines right through the film emulsion and diffuses and reflects off the film pressure plate, exposing the film from the backside, resulting in a lovely soft glow around the really bright parts of the image. This gives images a dreamy, ethereal look and I love adding this to an image in Photoshop. Using a few steps in an “IR Glow” action I created, I get this…

Before getting into how I added the glow, I want to backtrack a little and detail how I got to the B&W version of the image. Starting with the image opened in Photoshop, before doing a channel swap, I apply the following Black and White conversion settings…

Those settings make the reds and oranges of the original image a lot darker, adding to the dramatic B&W contrast that I like. I also add some more overall contrast using curves, and that results in this…

With the added glow, I then finally get this…

So, how do I add the glow? It is an action which does the following things…

1. Use Color Range to select the brightest parts of the image.
2. Copy those selected highlights to a new layer. Desaturate the highlights if being used on a colour version of the image to avoid over-the-top colour.
3. Crank the contrast on the new layer using a curve to make the brightest parts even brighter.
4. Use Gaussian Blur (16 pixels for these images) to spread out the highlights, which will then create that halo around the highlights. The blur radius can be adjusted to taste. Since the X-E2 is only a 16MP camera, the pixel radius is quite small, however if you are using a high-res camera, you’d want to increase that to between 20 and 30, maybe even higher.
5. Change the glow layer’s settings to Hard Light and 85% opacity.
6. Duplicate the glow layer and combine the two to make a more intense glow.
7. You can then play with the opacity to make the glow stronger or weaker and even play with various layer blending modes to change the effect.

Here are the results after step 3, with a black layer inserted below the glow layer to make it obvious. You can see that only the brightest highlights of the image have been duplicated for the glow layer…

Lastly, after the Gaussian Blur, and layer duplication, this is what one is left with (again, with just a black layer inserted below)…

One can get more sophisticated, which I have done myself as well, by selecting the bright areas and doing a smaller radius Gaussian Blur, then selecting only the very brightest highlights and using a larger blur radius to make those haloes a little larger around the brightest areas. Then, I’ve even reduced the opacity of the “interior” parts of the glow to maintain a little more detail and mostly restrict the glow to edges of bright objects. For the sake of brevity here though, I’ve kept it simpler since one can still get a nice result even with this fairly simple technique.

For the added grain layer, what I do is create a layer filled with 50% grey and set it to Overlay mode. A 50% grey overlay layer by itself has no effect on the image and one can then add grain to this layer independently of the main image to allow more flexibility in tweaking the look of the added grain. For the samples in the gallery, I’ve kept this fairly simple and only added monochrome gaussian noise, faded the noise to “darken” to avoid white speckles, which can be distracting and overwhelming, then scaled the layer to 125% in order to enlarge the grain and soften it slightly. I used the same process for the colour versions as well. Again, one can get very sophisticated with this and using selections based on the main image below, and add grain in differing amounts to the shadows, midtowns and highlights, to better mimic how film grain behaves. I didn’t bother with any screenshots of the added grain since it can only really be seen on the full-size images.

Note that some of the images in the gallery have had some tweaks to their exposure and opacity of the glow layer, although most tweaks were quite subtle after running the actions I created. Most images were processed identically using just the recorded actions of mine, although the samples in the article here are ever so slightly different than the final images in the gallery. As mentioned, you can do a lot of different things in post processing, so for example, if you’d like to mimic the trademark pink foliage one could get with Kodak’s long defunct Aerochrome EIR colour infrared film, well here you go…

Yes I know (for those in the know), Kodak EIR did have an anti-halation layer, so it technically wouldn’t have the glow I added to the above image, but what the heck, I like how it looks!

So that’s it for this tutorial. As always, if you have any questions about any of this processing, or about IR converted cameras in general, please feel free to drop me a line! In the meantime, enjoy the small image gallery linked again below!

Link to the image gallery with twelve IR image pairs: Colour and B&W Digital Infrared Gallery